What’s new?
New AI models across image, video, and 3D – You can now generate content using Seedance 1.5 Pro (image), grok-imagine-video (video), PixVerse V5.5 & PixVerse V5.5 Fast (video), and Hunyuan3D v3.1 & Hunyuan3D v3.1 Rapid (3D), so you have more creative flexibility across formats without switching tools.
Adjust similarity node in workflows– You can now fine-tune how closely outputs match reference images, so you get better creative consistency without over-constraining results.
Improved inference forms in Forge – You can now run inference with smarter, state-aware forms, so complex workflows are faster and require less reconfiguration.
Workflows productivity shortcuts – You can now duplicate nodes using keyboard shortcuts, so building and iterating on workflows is significantly faster. Users can now duplicate nodes using
CMD+D (Mac) or CTRL+D (Windows) for faster workflow building.
•Removed workflow details page - Workflow management has been streamlined by consolidating actions into the editor overview mode and updated workflow cards.
Bug fixes & improvements
Continuous improvements on the new unified export dialogue: Fixed export issues to ensure metadata is preserved, preferences are saved correctly, and exports behave reliably.
Fixed multiple issues affecting model training visibility, inference errors, and workflows stability.
Improved reliability across video models and inference workflows.